112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515 | class CouchbaseVectorStore(BasePydanticVectorStore):
"""Couchbase向量存储。
要使用，您应该已安装``couchbase`` python包。"""
stores_text: bool = True
flat_metadata: bool = True
# Default batch size
DEFAULT_BATCH_SIZE: int = 100
_cluster: Any = PrivateAttr()
_bucket: Any = PrivateAttr()
_scope: Any = PrivateAttr()
_collection: Any = PrivateAttr()
_bucket_name: str = PrivateAttr()
_scope_name: str = PrivateAttr()
_collection_name: str = PrivateAttr()
_index_name: str = PrivateAttr()
_id_key: str = PrivateAttr()
_text_key: str = PrivateAttr()
_embedding_key: str = PrivateAttr()
_metadata_key: str = PrivateAttr()
_scoped_index: bool = PrivateAttr()
def __init__(
self,
cluster: Any,
bucket_name: str,
scope_name: str,
collection_name: str,
index_name: str,
text_key: Optional[str] = "text",
embedding_key: Optional[str] = "embedding",
metadata_key: Optional[str] = "metadata",
scoped_index: bool = True,
) -> None:
"""初始化到Couchbase Vector Store的连接。
Args:
cluster (Cluster): 带有活动连接的Couchbase集群对象。
bucket_name (str): 要存储文档的桶的名称。
scope_name (str): 要存储文档的桶中的作用域的名称。
collection_name (str): 要存储文档的作用域中的集合的名称。
index_name (str): 搜索索引的名称。
text_key (Optional[str], optional): 文档文本的字段。默认为"text"。
embedding_key (Optional[str], optional): 文档嵌入的字段。默认为"embedding"。
metadata_key (Optional[str], optional): 文档元数据的字段。默认为"metadata"。
scoped_index (Optional[bool]): 指定索引是否为作用域索引。默认设置为True。
Returns:
None
"""
try:
from couchbase.cluster import Cluster
except ImportError as e:
raise ImportError(
"Could not import couchbase python package. "
"Please install couchbase SDK with `pip install couchbase`."
)
if not isinstance(cluster, Cluster):
raise ValueError(
f"cluster should be an instance of couchbase.Cluster, "
f"got {type(cluster)}"
)
self._cluster = cluster
if not bucket_name:
raise ValueError("bucket_name must be provided.")
if not scope_name:
raise ValueError("scope_name must be provided.")
if not collection_name:
raise ValueError("collection_name must be provided.")
if not index_name:
raise ValueError("index_name must be provided.")
self._bucket_name = bucket_name
self._scope_name = scope_name
self._collection_name = collection_name
self._text_key = text_key
self._embedding_key = embedding_key
self._index_name = index_name
self._metadata_key = metadata_key
self._scoped_index = scoped_index
# Check if the bucket exists
if not self._check_bucket_exists():
raise ValueError(
f"Bucket {self._bucket_name} does not exist. "
" Please create the bucket before searching."
)
try:
self._bucket = self._cluster.bucket(self._bucket_name)
self._scope = self._bucket.scope(self._scope_name)
self._collection = self._scope.collection(self._collection_name)
except Exception as e:
raise ValueError(
"Error connecting to couchbase. "
"Please check the connection and credentials."
) from e
# Check if the scope and collection exists. Throws ValueError if they don't
try:
self._check_scope_and_collection_exists()
except Exception as e:
raise
# Check if the index exists. Throws ValueError if it doesn't
try:
self._check_index_exists()
except Exception as e:
raise
self._bucket = self._cluster.bucket(self._bucket_name)
self._scope = self._bucket.scope(self._scope_name)
self._collection = self._scope.collection(self._collection_name)
super().__init__()
def add(self, nodes: List[BaseNode], **kwargs: Any) -> List[str]:
"""将节点添加到集合中并返回它们的文档ID。
Args:
nodes（List[BaseNode]）：要添加的节点列表。
**kwargs（Any）：额外的关键字参数。
batch_size（int）：批量插入的批量大小。
Returns:
List[str]：添加节点的文档ID列表。
"""
from couchbase.exceptions import DocumentExistsException
batch_size = kwargs.get("batch_size", self.DEFAULT_BATCH_SIZE)
documents_to_insert = []
doc_ids = []
for node in nodes:
metadata = node_to_metadata_dict(
node,
remove_text=True,
text_field=self._text_key,
flat_metadata=self.flat_metadata,
)
doc_id: str = node.node_id
doc = {
self._text_key: node.get_content(metadata_mode=MetadataMode.NONE),
self._embedding_key: node.embedding,
self._metadata_key: metadata,
}
documents_to_insert.append({doc_id: doc})
for i in range(0, len(documents_to_insert), batch_size):
batch = documents_to_insert[i : i + batch_size]
try:
# convert the list of dicts to a single dict for batch insert
insert_batch = {}
for doc in batch:
insert_batch.update(doc)
logger.debug("Inserting batch of documents to Couchbase", insert_batch)
# upsert the batch of documents into the collection
result = self._collection.upsert_multi(insert_batch)
logger.debug(f"Insert result: {result.all_ok}")
if result.all_ok:
doc_ids.extend(insert_batch.keys())
except DocumentExistsException as e:
logger.debug(f"Document already exists: {e}")
logger.debug("Inserted batch of documents to Couchbase")
return doc_ids
def delete(self, ref_doc_id: str, **kwargs: Any) -> None:
"""根据参考文档ID删除文档。
Args:
ref_doc_id：要删除的参考文档ID。
Returns:
无
"""
try:
document_field = self._metadata_key + ".ref_doc_id"
self._scope.query(
f"DELETE FROM `{self._collection_name}` WHERE {document_field} = '{ref_doc_id}'"
).execute()
logger.debug(f"Deleted document {ref_doc_id}")
except Exception:
logger.error(f"Error deleting document {ref_doc_id}")
raise
def query(self, query: VectorStoreQuery, **kwargs: Any) -> VectorStoreQueryResult:
"""在向量存储中执行查询并返回结果。
Args:
query（VectorStoreQuery）：包含搜索参数的查询对象。
**kwargs（Any）：额外的关键字参数。
cb_search_options（Dict）：传递给Couchbase搜索的搜索选项。
Returns:
VectorStoreQueryResult：包含前k个节点、相似度和id的查询结果。
"""
import couchbase.search as search
from couchbase.options import SearchOptions
from couchbase.vector_search import VectorQuery, VectorSearch
fields = query.output_fields
if not fields:
fields = ["*"]
# Document text field needs to be returned from the search
if self._text_key not in fields and fields != ["*"]:
fields.append(self._text_key)
logger.debug("Output Fields: ", fields)
k = query.similarity_top_k
# Get the search options
search_options = kwargs.get("cb_search_options", {})
if search_options and query.filters:
raise ValueError("Cannot use both filters and cb_search_options")
elif query.filters:
couchbase_options = _to_couchbase_filter(query.filters)
logger.debug(f"Filters transformed to Couchbase: {couchbase_options}")
search_options = couchbase_options
logger.debug(f"Filters: {search_options}")
# Create Search Request
search_req = search.SearchRequest.create(
VectorSearch.from_vector_query(
VectorQuery(
self._embedding_key,
query.query_embedding,
k,
)
)
)
try:
logger.debug("Querying Couchbase")
if self._scoped_index:
search_iter = self._scope.search(
self._index_name,
search_req,
SearchOptions(limit=k, fields=fields, raw=search_options),
)
else:
search_iter = self._cluster.search(
self._index_name,
search_req,
SearchOptions(limit=k, fields=fields, raw=search_options),
)
except Exception as e:
logger.debug(f"Search failed with error {e}")
raise ValueError(f"Search failed with error: {e}")
top_k_nodes = []
top_k_scores = []
top_k_ids = []
# Parse the results
for result in search_iter.rows():
text = result.fields.pop(self._text_key, "")
score = result.score
# Format the metadata into a dictionary
metadata_dict = self._format_metadata(result.fields)
id = result.id
try:
node = metadata_dict_to_node(metadata_dict, text)
except Exception:
# Deprecated legacy logic for backwards compatibility
node = TextNode(
text=text,
id_=id,
score=score,
metadata=metadata_dict,
)
top_k_nodes.append(node)
top_k_scores.append(score)
top_k_ids.append(id)
return VectorStoreQueryResult(
nodes=top_k_nodes, similarities=top_k_scores, ids=top_k_ids
)
@property
def client(self) -> Any:
"""
属性函数，用于访问客户端属性。
"""
return self._cluster
def _check_bucket_exists(self) -> bool:
"""检查链接的Couchbase集群中是否存在存储桶。
返回：
如果存储桶存在则返回True
"""
bucket_manager = self._cluster.buckets()
try:
bucket_manager.get_bucket(self._bucket_name)
return True
except Exception as e:
logger.debug("Error checking if bucket exists:", e)
return False
def _check_scope_and_collection_exists(self) -> bool:
"""检查链接的Couchbase存储桶中是否存在作用域和集合
返回：
如果存储桶中存在作用域和集合，则返回True
如果未找到作用域或集合，则引发ValueError异常。
"""
scope_collection_map: Dict[str, Any] = {}
# Get a list of all scopes in the bucket
for scope in self._bucket.collections().get_all_scopes():
scope_collection_map[scope.name] = []
# Get a list of all the collections in the scope
for collection in scope.collections:
scope_collection_map[scope.name].append(collection.name)
# Check if the scope exists
if self._scope_name not in scope_collection_map:
raise ValueError(
f"Scope {self._scope_name} not found in Couchbase "
f"bucket {self._bucket_name}"
)
# Check if the collection exists in the scope
if self._collection_name not in scope_collection_map[self._scope_name]:
raise ValueError(
f"Collection {self._collection_name} not found in scope "
f"{self._scope_name} in Couchbase bucket {self._bucket_name}"
)
return True
def _check_index_exists(self) -> bool:
"""检查链接的Couchbase集群中是否存在搜索索引
返回：
bool: 如果索引存在则为True，否则为False。
如果索引不存在则引发ValueError。
"""
if self._scoped_index:
all_indexes = [
index.name for index in self._scope.search_indexes().get_all_indexes()
]
if self._index_name not in all_indexes:
raise ValueError(
f"Index {self._index_name} does not exist. "
" Please create the index before searching."
)
else:
all_indexes = [
index.name for index in self._cluster.search_indexes().get_all_indexes()
]
if self._index_name not in all_indexes:
raise ValueError(
f"Index {self._index_name} does not exist. "
" Please create the index before searching."
)
return True
def _format_metadata(self, row_fields: Dict[str, Any]) -> Dict[str, Any]:
"""辅助方法，用于格式化Couchbase搜索API的元数据。
Args:
row_fields（Dict[str，Any]）：要格式化的字段。
Returns:
Dict[str，Any]：格式化后的元数据。
"""
metadata = {}
for key, value in row_fields.items():
# Couchbase Search returns the metadata key with a prefix
# `metadata.` We remove it to get the original metadata key
if key.startswith(self._metadata_key):
new_key = key.split(self._metadata_key + ".")[-1]
metadata[new_key] = value
else:
metadata[key] = value
return metadata
|